Is the road to machine learning paved with API providers?
- 12 February, 2016 12:10
When it comes to machine learning, the future is already here, but it’s not yet evenly distributed. Taking advantage of breakthroughs in the field can require a lot of work, which is tough for small companies and those without a whole team to build custom applications and algorithms.
According to Okta CEO Todd McKinnon, there’s a lot of hype around the potential of machine learning, but companies aren’t actually taking advantage of it. It's similar to how people discussed big data a few years ago.
“We think about this a lot, and the most interesting thing about machine learning that I’ve noticed over the last year is that it’s kind of like what big data was three years ago,” he said. "Everyone talks about it, but nobody really has it."
Mixpanel CEO Suhail Doshi sees that as a solvable problem. In his view, tech companies need to create and sell intelligent services that let other businesses use machine learning to perform key tasks.
In his view, a very small number of companies will have the people necessary to develop new machine-learning techniques, and only slightly more will have the team to help them get insights out of their data.
“Most companies don’t have the data scientists that are going to find signal from noise, so really what they have is data, at the end of the day,” Doshi said during a panel discussion in San Francisco on Thursday. "People are going to have to find a way to give up that data and say, ‘Here’s all my data, tell me something interesting about this.’"
Doshi's predictions line up with others from around the tech industry. Microsoft’s data chief, Joseph Sirosh, has said he expects to see a marketplace of intelligent algorithms and applications that companies can buy.
Microsoft already offers that through its Cortana Analytics Suite and Project Oxford tools, while IBM offers its Watson artificial intelligence APIs as cloud services. Algorithmia is a Seattle-based startup also creating such a marketplace.
But it’s not all fun and games, especially for an enterprise, according to Cumulus Networks CEO JR Rivers. In his view, laws governing companies, especially large ones, will prevent them from sharing data with providers of those services.
“Data privacy gets in the way of data transfer from any one company to another,” Rivers said. "So there’s no way any one company can give Google stuff with faces in it and let Google come back with answers — a real company, a big company. It’s not going to happen."
Then there’s the matter of what happens to applications powered by a particular service when its provider goes out of business. Right now, developers are rushing to reconfigure applications that rely on Parse, a set of services that Facebook sold for a couple of years, then announced it would be shutting down in 2017.
Parse’s exit is probably the best-case scenario for developers. They have a long time to update their apps, and Facebook has decided to open-source Parse Server software, so developers can run their own Parse instance if they need to.
Companies won't always be so lucky, though. It’s something to contemplate as we head toward this new API marketplace future.